MOVING FORWARD IN INDIA
An Analysis of IKEA’s Indian Distribution Network Development Master’s Thesis in Logistics & Transport Management
Gunnar Löfstedt
Supervisor: Johan Woxenius Graduate School
Date: 27 May, 2018
1
Abstract
Warehouses and distribution centers form an integral part of the supply chain. However, determining how many are needed to effectively supply a market is a complex question. This is especially true when entering a new market. IKEA has long been a company heavily focused on efficient logistics operations and is set to enter the Indian market in November 2018. This thesis identifies the main differences impacting distribution between the European and Indian market and investigates how these differences can impact IKEA’s distribution strategy in India.
To conduct this case study, several in-depth interviews with IKEA personnel based in both Europe and India were conducted and their answers were contrasted to the existing literature base. Additional comparisons were also made with another interviewee from the private freight transport sector in India. The findings revealed that while there are similarities between the European and Indian market in terms of geographical size, potential local sourcing and ownership structure, there are significant differences in terms of available infrastructure, land acquisition, and state legislations. The result is that deliveries in India are much less reliable from a time standpoint which would suggest the need to pursue a decentralized distribution strategy.
2
Acknowledgements
First and foremost, I would like to thank my supervisor Johan Woxenius for his support and feedback throughout the thesis. Secondly, I would like to thank my former manager Anders Lindqvist for providing help and inspiration in this project and for the recommendations of whom to contact for interviews. Special thanks are also extended to the interviewees who partook in this project and provided invaluable insights. Without their aid, the project could not have been completed.
3
Abbreviations
3PL – Third Party Logistics Provider CDC – Central Distribution Center DC – Distribution Center
DDC – Decentralized Distribution Center FDI – Foreign Direct Investment
ICD – Inland Container Depot MNC – Multinational Corporation NDC – National Distribution Center RDC – Regional Distribution Center
4
Table of Contents
Abstract ... 1
Acknowledgements ... 2
Abbreviations ... 3
1. Introduction ... 1
1.1 Background ... 1
1.2 Case Company Description ... 2
1.3 Problem ... 3
1.4 Purpose ... 3
1.5 Research Gap ... 3
1.6 Research Question ... 3
1.7 Scope ... 4
2. Methodology... 4
2.1 Research Approach ... 4
2.2 Scientific Philosophy ... 5
2.3 Scientific Approach ... 5
2.4 Research Strategy ... 6
2.5 Data Collection ... 6
2.5.1 Case selection ... 6
2.5.2 Interview Method ... 7
2.5.3 Secondary Data ... 8
2.6 Method for Data Analysis ... 8
2.7 Research Trustworthiness ... 9
2.7.1 Reliability ... 9
2.7.2 Validity ... 10
2.8 Ethical Implications ... 11
3. Theoretical Framework ... 12
3.1 Warehousing ... 12
3.2 Decentralized vs Centralized Distribution ... 13
3.2.1 Decentralized Distribution ... 13
3.2.2 Centralized Distribution ... 15
3.2.3 Expanding the Literature Base ... 16
4. Factors Impacting Location Selection & Number of DCs ... 19
4.1 Lumsden’s Framework ... 19
5
4.2 Issues of Land Acquisition ... 20
4.3 Ownership Type ... 23
4.3.1 Public Warehouses ... 24
4.3.2 Private Warehouses ... 25
4.3.3 Third Party Logistics Providers (3PLs) ... 26
4.4 Available Infrastructure ... 27
4.4.1 Maritime Connectivity... 27
4.4.2 Rail Connectivity ... 29
4.4.3 Road Connectivity ... 30
5. Findings ... 33
5.1 IKEA’s European Distribution Network ... 33
5.2 Issues of Land Acquisition & Ownership Type ... 35
5.3 Available Infrastructure ... 36
5.3.1 Maritime Connectivity... 36
5.3.2 Rail Connectivity ... 36
5.3.3 Road Connectivity ... 37
5.4 IKEA’s Distribution Network in India ... 37
5.5 Issues of Land Acquisition & Ownership type ... 37
5.6 Available Infrastructure ... 40
5.6.1 Maritime Connectivity & Customs Clearance ... 40
5.6.2 Rail Connectivity ... 40
5.6.3 Road Connectivity ... 42
6. Analysis ... 44
6.1 Issues of Land Acquisition and Ownership Type ... 44
6.2 Available Infrastructure ... 45
6.3 Number of DCs and Size of DCs in Europe ... 50
6.4 Number of DCs and Size of DCs in India ... 51
7. Conclusion ... 54
8. Delimitations & Further Research ... 57
9. References ... 58
10. Appendix ... 65
10.1 Interview Guide IKEA Personnel: ... 65
10.2 Interview Guide Trucking Company ... 66
6
List of Figures
Figure 1 Transport Distances in Centralized vs Decentralized distribution. ... 14
Figure 2. No of Warehouses and Costs. ... 17
Figure 3. Local DCs and RDCs Europe.. ... 18
Figure 4. Local DCs and RDC USA. Source: ... 18
Figure 5. Framework of Factors Impacting Warehouses and DC location.. ... 19
Figure 6 Land Prices in Europe. ... 22
Figure 7. Road Network in India.. ... 32
Figure 8 IKEAs European DCs. ... 34
Figure 9. Poorly Maintained Warehouse. ... 39
Figure 10. Warehouse with ground level loading. ... 39
Figure 11. A Rail Terminal with a Large Unfenced Area.. ... 41
Figure 12. Sign Located at Rail Terminal Office. ... 42
Figure 13 Road network in India. ... 50
Figure 14 No of Warehouses and Costs.. ... 52
1
1. Introduction
In this chapter, the general background and the case company will be introduced. The problem, purpose, and research gap along with the research question will also be presented.
1.1 Background
Warehouses or distribution centers (DCs) have long formed an essential part of a supply chain.
Over the years, the size and functions of DCs have changed significantly. In the early 20th century, most DCs were smaller warehouses which stored finished goods. Today, they take a wide variety of forms depending on the type of goods flowing through, the region in which they exist, and purpose. However, in addition to the function they fill, there are also a wide array of differences between different countries. Though there is a substantial amount of work dedicated to the development of distribution networks, with particular focus on the European and US context, the amount of literature focusing on the developing nations, such as India, is limited (Ng & Cetin. 2012).
Ever since India joined the World Trade Organization (WTO) in 1995, it has gradually liberalized trade legislation to attract more Foreign Direct Investment (FDI). Naturally, with its large domestic market abundance of cheap labor many have been viewing India as the next market to invest in. However, while growth rates have remained fairly high since the early 1990s, they have not reached the levels seen by countries such as China. As a result, the total GDP of India has remained comparatively low (Dash & Sharma, 2011). There are several explanations about why GDP growth has been slower than what many hoped for. Despite opening its market to FDI, foreign companies have faced many challenges in India. One of the most common challenges relates to the poor state of the national infrastructure. Indeed the positive correlation between infrastructure development and GDP growth has been established and confirmed by several authors including Simon & Natarajan (2017). The challenges related to the state of the infrastructure undeniably impacts the logistics costs.
However, there are also other challenges which stem partly from the sheer size of India. Due to the numerous states of India there are a wide array of different laws which impact business operations. Previously, each state had their own form of taxation which led to a very complex tax structure for companies operating in several states. Though this has been addressed with new legislation and introduction of the Goods and Services Tax (GST) in 2016, other hurdles remain, especially in regard to infrastructure needs and land acquisition (Haralambide & Gujar, 2011. Deloitte, 2014. Sunitha & Chandra, 2017). All these factors have a direct impact on the
2 distribution strategies of companies. These challenges will likely also be faced by the Swedish furniture retailer IKEA which is in the process of establishing itself in the Indian market, with its first store opening in Hyderabad later in 2018.
1.2 Case Company Description
Efficient logistics operations have been at the heart of IKEA’s operations ever since it was founded in 1943. This logistics focus eventually led to the development of one of the most defining characteristics of the company, the unassembled furniture in space reducing flat packaging (Klevås, 2005. Wever, 2011). This focus has also been mirrored in its use of transport. IKEA has a history of utilizing intermodal transport in Europe primarily in efforts to reduce transport pollution. Though IKEA’s own rail service was discontinued in 2004, intermodal transport is still utilized to this day in combination with long distance trucks (Roso, Woxenius & Olandersson, 2006). Early on, IKEA was also able to utilize the benefits of international sourcing. With the Swedish suppliers primarily supplying competing furniture retailers, IKEA expanded its sourcing base abroad to countries like Poland which was geographically close and had more affordable labor. These factors would eventually form an integral part of IKEA’s strategy as they expanded. Today IKEA is the world’s largest furniture retailer with 403 stores in 49 countries. 275 of these are located in Europe, 56 are located in North America and 47 stores are in Asia (Inter IKEA Group, 2017). The company also has approximately 1220 suppliers in 55 countries including Poland, Sweden, France, Russia and China (IKEA, Purchasing, 2018. IKEA Industry, 2018). Though Europe accounted for approximately 81% of IKEA’s turnover in the early 2000s, other markets have grown over the years (Johansson & Thelander, 2009). As of 2018, Europe accounted for approximately 75%
of the market.
The global supplier network means that in order to be able to deliver the products to the stores, there needs to be an efficient distribution network. After the first Indian IKEA store opens in 2018, the aim is to open 25 stores in India by 2025 with Mumbai, Bangalore, New Delhi Chennai, and Gurgaon earmarked as primary areas (Govind, 2017. Jain, 2017. Financial Express, Nov, 2, 2017). Currently much of the distribution network is only in the planning stages. However, IKEA already has several suppliers in India with whom they have collaborated with over the past 30 years. These connections will likely play an increasingly important role as IKEA expands production in India to meet the 30% required production volume established by the Indian government. Meanwhile, the remaining 70% will still be imported from abroad (IKEA India, 2018. Malviya, 2018). Naturally there are a wide array of
3 challenges that require attention. With the aim to rapidly expand their position in the Indian market, the distribution network needs to be able to facilitate timely deliveries to the stores.
This is particularly challenging in a market notorious for the poor state of its infrastructure as well as multiple other challenges that can stall progress.
1.3 Problem
In recent years, the number of companies relying on centralized distribution centers (CDCs) has increased substantially. The main reason for pursuing such a strategy is to cut costs (McKinnon, 2009). However the downside with this strategy is that companies are instead increasingly reliant on one location and good connectivity. Whereas this may not be too problematic in North America or Europe this may not be true for other parts of the world. With its growing economy, massive population and low labor costs, many companies such as IKEA are eyeing India as the next market to enter. However, India is renowned for the poor state of its infrastructure and high amount of bureaucracy (McKinsey & Company, 2010. Deloitte, 2014). Since IKEA is a logistics focused company, these challenges are particularly relevant as they will have a direct impact on the effectivity of IKEA’s distribution and warehousing.
1.4 Purpose
The purpose of this research is twofold. First and foremost, the aim is to determine to what extent the existing theoretical framework about distribution centers can be applied in an Indian setting and what the key differences are between India and the western hemisphere, primarily Europe. The aim of this paper is to expand on the current literature base by providing a case study about this topic. Secondly, the aim is to produce insights that can be utilized by IKEA when developing their operations in India.
1.5 Research Gap
Though plenty of research about distribution centers has been conducted such as that of McKinnon (2009) Cidell, (2010) and Rushton, Croucher, & Baker (2014) most of these studies have been conducted in either North America or Europe. The amount of research focusing on India is currently limited as noted by Ng & Cetin (2012).
1.6 Research Question
In order to appropriately study the phenomena, the following research question will be studied;
How can IKEA develop their distribution network in India compared to how it is in Europe?
4 In order to answer this question, there are several sub questions that need to be addressed. These are as follows:
1. Is there a notable shift in IKEA’s ownership structure between Europe and India?
2. How do local governments impact IKEA’s warehouse location selection in Europe and India respectively?
3. How do differences in infrastructure and freight transport in the European and Indian market affect IKEA?
4. How many distribution centers will IKEA use to serve the Indian market compared to the European market?
1.7 Scope
Since this is a single case study, the scope of this paper is limited to IKEA’s distribution in the European and Indian market. While the aim is to expand on the existing literature base by providing an in-depth case study with results that are applicable in other situations, it is acknowledged that certain findings may be specific for the case study company and the industry it operates in. Nonetheless, the factors analyzed should provide insights to the limitations of the Indian market and aid future research within the academic and professional communities. It should also be acknowledged that this thesis is using the perspective of a western company entering the Indian market. As such, there are certain limitations in the scope. In addition, the intention of this study is not to create new theories but merely apply existing theories in a different geographical context.
2. Methodology
In this chapter, the methodology used to produce this thesis will be introduced and explained.
Thereafter, the reasoning behind the method selection will be examined and further elaborated.
2.1 Research Approach
First and foremost, it has been argued that a researcher should clearly define what research approach will be applied as this impacts the type of data that will be collected. The two dominant options to decide between are qualitative or quantitative studies. Qualitative studies rely on non-numerical data as sources for information whereas quantitative studies rely primarily on numerical data as sources (Saunders, Lewis, & Thornhill, 2009). Since this thesis is aimed at comparing the distribution strategy of IKEA to existing literature, the qualitative research method was selected. It was also argued that the qualitative method is best suited for
5 researchers attempting to create or expand on existing theories from the data and information gathered (Saunders, et al. 2009).
2.2 Scientific Philosophy
Among researchers, there are three primary scientific philosophies: positivism, interpretivism, and hermeneutics. As positivism is closely linked to mathematical analysis and inferential statistics and is primarily used within the natural sciences, it is not particularly suitable for this research (Lee, 1991. Saunders, et al. 2009). Interpretivism focuses more on the relations between different people and the role that each of them play (Saunders, et al. 2009). In contrast, hermeneutics focuses entirely on textual analysis and is most commonly used among researchers studying differences in language (Butler, 1998). Therefore, it is not particularly suitable for those studying the operations of a company.
Furthermore, those who use the interpretivistic philosophy typically rely on fewer, but more in- depth interviews than many of the other philosophies (Saunders, et al. 2009). Due to the qualitative nature of this study this philosophy was considered to be the most suitable.
2.3 Scientific Approach
There are three main scientific approaches. These are identified as, inductive, deductive, and abductive (Dubois & Gadde, 2002. Saunders, et al. 2009). An inductive study is made by first gathering information and thereafter tying these findings to a theory (Locke, 2007). This approach of case study methodology was also referred to as Theory Generation by Ketokivi &
Choi (2014). As indicated by the name, this approach is best suited for researchers attempting to produce new theories. Since this is not the purpose of this thesis, the inductive approach was not chosen.
In a deductive study, on the other hand, researchers begin with a model or theory as a base which the observations and findings are compared to. The deductive approach has been criticized due to the limited amount of suggestions about how that type of research would be conducted (Barratt, Choi, & Li, 2011). However, other authors have argued that the deductive approach is highly suitable for qualitative case studies when the point is to either confirm or falsify an existing theory (Locke, 2007).
An abductive study utilizes systematic combining which is a process of continuously combining the theoretical framework and the empirical information in order to produce a hypothesis (Baral, 2000. Dubois & Gadde, 2002). The abductive approach has been argued to be particularly suitable for research that utilizes case studies as a research strategy due to the unique
6 opportunities that case studies provides researchers to develop theories (Dubois & Gadde, 2002). It has also been argued that the abductive approach is best suited if the purpose is to use an existing theory and expand on it (Ketokivi & Choi, 2014).
Since this thesis investigates IKEA and their operations in India, this thesis will utilize the case study research strategy, which will be explained in greater depth in Chapter 2.4. Since the purpose is to apply a set of theoretical assumptions in a new context, here India, the abductive approach was found to be most useful.
2.4 Research Strategy
Since this paper aims to compare a theoretical framework with the findings from a company the case study research strategy was selected. While this strategy has been criticized by several authors for lacking a clear approach and guidelines of how to apply the strategy, is has also been argued that since each case is unique, a clear set of guidelines that would be applicable in all circumstances would be difficult to develop (Eisenhardt, 1989. Stake, 1995). In order to address this issue, Stake (1995) developed three categories of case studies. These were defined as: instrumental, intrinsic, or collective case studies. An instrumental case study is defined as a case study in which a case company is studied in order to produce insights that are applicable in several contexts. This is in contrast to the intrinsic case study which only aims to produce insights that are applicable in that particular context. A collective case study, on the other hand, is a case study that has several focus points, e.g. multiple companies or individuals. While this paper utilizes information from numerous sources and interviewees, there is but one company that serves as the focus of the thesis the aim of which is to produce insights which are applicable to other companies. As such, this case study is an instrumental case study.
2.5 Data Collection 2.5.1 Case selection
One of the key requirements of an instrumental case study is that it can provide insight into a particular issue. As previously mentioned, IKEA is an international Multinational Corporation (MNC) which is in the process of establishing itself in the Indian market. Many other western MNCs are also attempting to establish themselves in the Indian market due to its growing economic strength and consumer demand. In addition, IKEA is a highly relevant company to research due to its emphasis on efficient logistics operations, which is the focus of this thesis.
A key criteria mentioned by Stake (1995) relates to accessibility. In order to produce a case study, the author needs access to interviewees as well as other information from that company.
7 In this instance, the reason why IKEA was selected as a case study company was due to the researcher’s previous work experience at there. Therefore, the researcher was in a unique position to utilize these previous experiences and contacts made to gather the relevant material, conduct the relevant interviews and produce this thesis. Additional information regarding logistics operations in India were gathered through a field trip to Bangalore, where IKEA’s Indian Corporate Offices are located. Face to Face interviews were conducted with both IKEAs Distribution Operations Manager and a board member of an Indian trucking company during this trip.
2.5.2 Interview Method
The two primary interview methods are labelled as either structured or semi-structured. The main difference between the two interview methods is that in a structured interview, the interviewee is not allowed to deviate from the topic whereas in a semi-structured interview, the interviewee is allowed to deviate and more freely expand on the questions asked (Saunders et al. 2009). In order to gain access to more of the insights that the interviewee may have, the semi-structured method was chosen.
A combination of face to face interviews, interviews via skype and follow up interviews via email were conducted. The interview guides can be found in the appendix, 10.1 and 10.2. The process of conducting interviews via email have been criticized due to the possibility of missing certain information since it is difficult to interpret tone or expression (Lee, 1994). Another critique is that the time lag between question and answers may result in unfocused answers and limits the possibility to ask follow up questions (James & Busher, 2006). This is especially relevant when the semi structured interview method is chosen. However, by combining face to face interviews and interviews via skype, some of these issues can be addressed. Though it will not be possible to observe body language when conducting email interviews, it has been argued that the tone of the interviewee can still be understood, especially if there have been face to face interviews previously (Lee, 1994).
Interviews were conducted with 4 interviewees. The interviewees were selected based on their familiarity with the topic and on the recommendations from other interviewees. This is referred to as networking or snowball sampling by Collis & Hussey (2013). An interviewee from outside IKEA was interviewed in order to gain additional viewpoints. This interviewee was selected based on their knowledge of the topic and previous associations with the author. The titles and the type of interview each interviewee participated in are outlined below:
8 Interviewee Name (as
denoted in thesis)
Position Number of Interviews and
Interview Method Interviewee 1 Distribution Operations Manager
- IKEA India
1 Interview via Phone 1 Interview via Email 1 Interview Face to Face Interviewee 2 Multi-Channel Network Design &
Implementation Manager -EU
1 Interview via Phone
Interviewee 3 Global / EU Network Distribution Design Manager
1 Interview via Skype
Interviewee 4 Member of company board
-Private trucking company in India
1 Interview Face to Face
Table 1. Summary of Interviews and Interviewees
2.5.3 Secondary Data
Lumsden’s framework (2006) of factors impacting warehouse and DC location served as the basis of what secondary information to gather. A wide array of academic articles were selected based on relevance, impact factor, and times referenced. These articles were complimented with several newspapers articles as well as case studies produced by several leading consulting firms.
These sources were used to obtain up-to-date information as well as in-depth studies about specific subjects for which there was a lack of academic research, such as a study comparing local attributes for warehousing across the largest cities of India.
The articles focused primarily on general infrastructure and infrastructure development in India and Europe. Issues related to land acquisition were also studied and a wide range of articles about warehouses & DCs and different types of ownership were studied as well. To find relevant articles, google scholar was used extensively in combination with course material gathered during graduate studies in Sweden and India.
2.6 Method for Data Analysis
Though there is no standard format to apply when analyzing information gathered in a case study there are several recommendations that researchers can apply. One of the methods Saunders, et al, (2009) recommend a researcher undertake is to write short summaries of the key topics. It was argued that this would help the researcher find similarities and differences between the different sources of information. This step should be complemented by transcribing the interviews in order to help highlight some of the key themes (Saunders, et al. 2009).
9 Similarly, Collis & Hussey (2013) suggested the following three steps when analyzing data (Pg.
157):
“Reducing the data
Displaying the data
Drawing conclusions and verifying the validity of those conclusions”
The first step of reducing the data involves simplifying and restructuring the data that has been gathered. This can be done by “using a pre-existing theoretical framework or one that emerges during the data collection stage to provide categories into which the data can be fitted” (Collis
& Hussey, 2013. Pg. 159). Once the interviews were transcribed, fitting the information into the relevant categories introduced in the literature review was straightforward. Using similar categories also makes it easier to display the data in a manner that is easy for the reader to follow. Likewise, Eisenhardt (1989) argued that once a sufficient amount of primary information has been gathered from the case company and secondary information from literature, the researcher should study them until distinguishable similarities appear.
2.7 Research Trustworthiness 2.7.1 Reliability
For research to be deemed trustworthy and of scientific value, it needs to be both reliable and valid. A study which is not reliable can, by definition, not be valid. In order to produce a reliable study it is recommended to answer the following three questions (Easterby-Smith et al. 2008.
Pg. 109. In Saunders, et al. 2009. Pg. 156):
1. “Will the measures yield the same results on other occasions?
2. Will similar observations be reached by other observers?
3. Is there transparency in how sense was made from the raw data?”
In relation to the first question, it is important to consider the interviewees that are selected. A particular employee may have very different opinions and answers from their colleagues and may respond differently depending on the circumstances. The time of the day and the day of the week can all impact the interviewee’s mood and response (Saunders, et al. 2009). One of the easiest ways to tackle this issue is to conduct several interviews at different times with the same interviewees.
Since this study is conducted by a single researcher, questions regarding bias must be addressed.
This is referred to as observer bias and relates to Question 2. One way to reduce observer bias
10 is to have other colleagues read through the questionnaire before the interview takes place and compare notes after the interview. Though this is a time consuming process it can help uncover potential faults or sources of bias for the researcher. Similarly having colleagues read through the study will also help in determining whether or not it is clear how the researcher understood and interpreted the raw data, as described in Question 3.
2.7.2 Validity
In terms of validity there are two types, namely, internal and external validity. Internal validity can be defined as: “the extent to which the findings can be attributed to the interventions rather than any flaws in your research design” (Saunders, et al. 2009. Pg 143). For qualitative studies, internal validity is directly impacted by the questionnaires produced for gathering primary data.
Here, particular emphasis is placed on whether or not the interviewee understands the question to its fullest extent and whether or not the interviewer understands the information provided by the interviewee (Saunders, et al. 2009). The second aspect that must be considered when producing a questionnaire is that the answers will measure what the researcher set out to measure. One way of tackling this issue is to have several preparatory interviews with colleagues from both inside and outside an academic setting to determine if the questions are adequately comprehensible. In addition, the interviewees of this research also received copies of the paper before the publication to read and comment on. This is referred to as respondent validation by Gibbs (2007) and was undertaken to ensure that the answers provided by the interviewees were correctly understood by the researcher. There are also additional benefits to utilize respondent validation. If the interviewee was mistaken about certain information provided during the interview, they would have the chance to clarify and correct themselves.
This is particularly important as the interviewer may have correctly interpreted the information provided by the interviewee but if the information provided was incorrect, it could impact the validity of the research (Gibbs, 2007). In addition, utilizing respondent validation could also impact the reliability of the study. If an answer was provided due to a specific circumstance, such as the time of day impacting mood, as mentioned in 2.7.1, the interviewee would now have the chance to either reword their answer or retract certain parts if the answers were the result of the circumstance at the time rather than fact.
External validity is commonly also referred to as generalizability and refers to the extent to which the findings produced in the study can be generalized for other studies. This is a particularly important aspect to consider when producing single case studies as the findings produced are based on a single company. However, proponents of the case study research
11 strategy have argued that it is not the sample in and of itself which is of primary concern but rather the findings produced (Barratt et al. 2011). Though it is not certain that the findings will be generalizable, this can be deduced by producing additional studies in the future (Saunders, et al. 2009).
2.8 Ethical Implications
Before conducting the study, there are multiple ethical considerations the researcher needs to make. In general it is recommended that the researcher: “should not subject those you are researching (the research population) to embarrassment, harm or any other material disadvantage” (Saunders, et al. 2009. P.g 160).
This also relates to the ethical issues raised by Blumberg, Cooper & Schindler (2005) on whether or not the findings in the research can be classified as good. The key argument was that research that produced more harm than good could not be justified. Due to the nature of this research, the primary costs would be related to the time the interviewees spent participating in the research as well as the potential material disadvantage. Since this is an in-depth case study about one company, there is a risk that the information produced herein could have been of classified nature. To avoid any material disadvantage, the interviewees were therefore allowed to read through the paper before publication. Though this raises the question of participant bias, this issue was tackled by discussing the contents in depth and negotiating which parts could be included and reformulated without being removed. Another point raised by Saunders, et al. (2009) relates to the issue of transparency. It has been argued that interviewees who are not informed of the purpose of the study may feel deceived or may even feel that the research can negatively impact them or their careers. Keeping true to this view all the interviewees were fully informed of the purpose of the study and the reasoning behind the interview questions. Furthermore, special care was taken to ensure anonymity for interviewees who requested it.
12
3. Theoretical Framework
In this chapter, the concepts of warehousing and distribution centers are explained and the two underlying inventory models in distribution networks are introduced.
3.1 Warehousing
The term warehouse has often been used interchangeably with the term Distribution Center (Higginson & Bookbinder, 2005. Rimiene & Grundey, 2007. Higgins, Ferguson & Kanaroglou 2012). However, there are some important distinctions to make. Even though the two terms are often used interchangeably there are several types of warehouses and the DC is just one of them, as Higginson & Bookbinder (2005) pointed out. Though this separation may seem trivial, it is an important distinction to make as early as possible before conducting further research due to the specific requirements each DC has. Higginson & Bookbinder (2005) divided up the modern DCs into seven categories according to the different key functions they filled. These were defined as: make bulk/break bulk consolidation center, cross-dock, transshipment facility, assembly facility, product fulfilment center, distribution center for returned goods and miscellaneous other roles. However, these separations can provide limited help at times since many DCs perform several functions, as Higginson & Bookbinder themselves pointed out (2005). Therefore, for the purpose of this study, it is more useful to rely on the definition utilized by Higgins, et al. (2012) which was initially developed by Notteboom & Rodrigue (2009). Here, they argued that DCs are logistics centers which provide 3 primary functions, (Higgins et al.
2012. Pg. 11):
1. “Transfers: the contents of maritime containers are transferred into domestic containers or truckloads.
2. Cross docking: the contents of inbound loads are sorted and transloaded to their final destinations.
3. Warehousing: a standard function still performed by a majority of distribution centers that act as buffers and points of consolidation and deconsolidation in supply chains”
This definition also closely matches that of Higginson & Bookbinder (2005) who stated that DCs are a supply chain node that serve to facilitate “storage of intermediate or finished goods;
consolidation of orders; and transportation” (Pg. 43). DCs can be a variety of sizes depending on the products and the volume they have to handle. Naturally, the size of the market or the amount of sales impacts the volume that a DC has to handle, however, the distribution strategy utilized also has an impact. This will be studied in depth in the upcoming section.
13
3.2 Decentralized vs Centralized Distribution
There are two dominant inventory models for a supply or distribution network; decentralized or centralized distribution. Companies that rely on multiple locations for their inventory needs use a decentralized inventory model whereas companies that rely on a single location for their inventory needs use a centralized inventory model (Schmitt, Sun, Snyder & Shen. 2015). These two models will be studied and explained in depth in the following section.
3.2.1 Decentralized Distribution
If supply risk is a larger issue than demand risk, many companies opt for a decentralized inventory model. In this model, several locations serve as storage or distribution points in order to reduce supply risk. The logic is that if suppliers are unreliable for any reason, the risk can be mitigated by having inventory at several locations. This is referred to decentralized distribution centers or DDC and is related to risk diversification (Snyder & Shen, 2006. Schmitt, et al. 2015).
For instance, if a region has two cities named City A and City B, and if the company supplying City A and City B were to have a high supplier risk, the company would be more likely to rely on multiple distribution locations, perhaps one in City A and one in City B, to decrease this risk. If there is a disruption in the supply to the warehouse in City A, the company will still most likely be able to serve City B as it has a separate warehouse (Schmitt, et al. 2015). Several factors impact supply risk such as production issues, weather (which may impact delivery times), administrative issues at the supplier’s end, border controls and availability of transport.
Generally speaking, the greater the distance between the supplier and DC or DC and the customers, the greater the supply risk (Zsidisin & Ellram, 2003). The term “supply risk” was also referred to as a company’s “anpassningförmåga” by Lumsden (2006) and can be translated into adaptability. Another potential advantage to having several DDCs that was emphasized by Lumsden (2006) is that it can decrease total transport distances. Indeed, if a very large country is served entirely by one DC, the total transport distances will be very high. This relationship is shown below:
14 Point of consumption (Store)
Distribution Center (DC)
Figure 1. Transport Distances in Centralized vs Decentralized distribution. Source: Lumsden, 2006. Pg. 623.
Note: Total Transportsträcka = Total transport distance
In this scenario, it should be noted that the transport distances are primarily shorter for outbound goods, i.e. goods moving from the DC to the store. If we consider inbound goods, one needs to take into consideration the source of production. If the DC is in a country with several production centers total transport distances can still be shorter. However, if there is one production center or one area where all the imported goods arrive, such as a port, the inbound transport distances may be larger, similar to the situation described in quadrant D in fig 1.
Therefore, it is important to find a balance between inbound transport distances and outbound transport distances.
Though most warehouses and DCs constructed prior to 1970s would have been classified as DDCs, the development of these has steadily declined since. The downside with DDCs is that overhead costs, as well as rent and employee costs are typically higher. Logically, the more DCs that are managed by the same company the more complex the organization becomes and the higher costs associated with coordination are. With many companies increasingly focused on cutting costs there was an increase in focus on developing CDCs (Chandra & Jain, 2007.
15 McKinnon, 2009. Rushton, et al. 2014). However, with the recent growth of e-commerce, there has been a renewed interest in developing more DDCs. There are two primary reasons for this.
The first is related to response times. Oftentimes e-commerce customers expect short deliver times which means that e-commerce companies have to have their DCs located near their customers (Lau, Jiang, Ip & Jiang, 2010). The second reason is in relation to transport costs.
Due to the nature of e-commerce, outgoing volumes are rather small and most of the parcels can be defined as low density. Each customer typically only buys one or a small handful of products each time meaning that economies of scale can be difficult to achieve even if each geographical area has several customers. Therefore, it stands to reason that many e-commerce retailers try to minimize the distance between the DC and the customers (Bendoly, Bretthauer
& Venkataramanan, 2007. Mattarocci & Pekdemir, 2017). McKinnon (2009) also argued that IT developments enable efficient management of inventory from a central location, even if the inventory is dispersed in several locations, which was previously a restriction for a decentralized distribution strategy. Nonetheless, it should still be noted that many companies in the e-commerce sector still utilize some type of decentralized distribution in combination with a more central location where products with lower turnover are typically located (Mattarocci &
Pekdemir, 2017).
3.2.2 Centralized Distribution
In order to reduce demand risk, many companies rely on centralized warehouses as key distribution points or CDC. The logic is that if there is uncertain demand in a larger region, costs can be saved by pooling the resources together (Schmitt, et al. 2015). This is often also referred to as risk-pooling (Snyder & Shen, 2006). Using city A and city B again as examples, if these cities have uncertain demand, it is often considered wise to rely on a central distribution location somewhere in-between these two cities so that the inventory can be stored at a single location. If city A were to see a large surge in demand for a certain product whereas city B does not experience that same surge in demand, the inventory at the single location can more easily be pooled together and shipped to city A to meet the demand. Later, if B experiences the same surge, but at a later date, ideally the stock levels will have been replenished and the inventory can be shipped to city B. Though it may sound simple, there are a wide array of factors which can impact demand risk or demand variation depending on the product sold. Any number of events can impact demand risk. Whereas some are more predictable, such as a change in seasons, others such as natural disasters, are more difficult to predict. Any such fluctuation in
16 demand can impact an organization negatively if there is a lack of extra inventory, or buffer (Zsidisin & Ellram, 2003).
The compared cost is also important to mention. Due to the reliance on a single location, it is far easier to achieve costs of scale for a CDC. This means that even if total transport distances increases, as suggested by Lumsden (2006) in Figure 1, there could still be cost savings due to fact that the vehicles have a higher fill rate or due to a potential modal shift in transport. Due to the higher volumes and longer distances, transport volumes could shift from road to e.g. rail, which is more competitive when distances and volumes are high. Other sources of increased efficiency in terms of costs include lower compared rent, lower overhead costs, decreased need for supervision, and fewer employees for the same volume (Rushton, et al. 2014). Such benefits are indeed attractive and have been mirrored by an increase in demand for larger warehouses.
In the UK, for instance, the amount of floor space grew an average of 2.5% between the years of 1998 and 2004 (McKinnon, 2009). Similar trends have also been witnessed in the US particularly since the late 1990s. As warehousing needs shifted from being production oriented to consumer oriented, the need for larger centralized storage locations increased (Cidell, 2010).
An additional factor that could help explain this trend is the fact that it is becoming increasingly common for smaller and midsize companies to outsource their warehousing needs to a third party logistics provider (3PL). It has been estimated that approximately 65% of the European DCs (EDC) are handled by a 3PL (De Koster & Warffemius, 2005). The impact this has is that these 3PLs typically have several customers and are thus more likely to use a single large facility to service all of them (Higginson & Bookbinder, 2005). Overall, it should be mentioned that CDCs are more common in mature markets as it is easier to determine the costs of inbound and outbound freight volumes and achieve costs of scale in transport as mentioned by Matarocci and Pekdemir (2017). While the use of CDCs present companies with a multitude of benefits, there are still certain drawbacks with this approach. Though land constraints are frequently among the top factors, an additional factor is the increased traffic volumes on a single, or very few routes as mentioned by Hesse & Rodrigue (2004).
3.2.3 Expanding the Literature Base
However, in most instances these two categories of DCs are insufficient to describe what companies rely on. While smaller companies may only rely on a single DC for inventory, this is rarely the case for larger companies and while many larger companies rely on several DCs, the number that are utilized vary greatly depending on the type of products, size of company and the distances between suppliers and end customers. As argued by Snyder & Shen (2006),
17 most companies experience both demand and supply uncertainty to a certain extent, and thus are likely to use some type of mix between CDC and DDCs. The relationship between number of warehouses and costs is perhaps best described by McKinnon in the image below:
As visible in the figure above, there is a relationship between cost and the number of warehouses. For most companies, the reliance on a single DC can bring about substantial costs due to the higher associated distances. Using the figure above, this would correspond to higher delivery costs as the distances between the customers or stores and the DC are larger, as shown previously in Figure 1 (Lumsden, 2006). However, in the other extreme, having an excessive amount of small DCs scattered throughout an area also has substantial costs as depicted in Figure 2. This is due to the fact that inventory costs in the large number of DCs would increase due to the higher overhead costs and number of employees required. If inventory costs were to increase substantially, this would cause the corresponding curve to rise, which would decrease the optimum number of warehouses. In contrast, if delivery costs were to increase instead, the optimum number of warehouses would increase. As McKinnon stated: “The centralisation of inventory, which has been a dominant logistics trend over the past 40 years, may be reversed by steep increases in the cost of freight transport and reductions in its speed and reliability.”
(2009. Pg 297). This reasoning is closely linked to Lumsden’s argument about transport distances (2006).
Figure 2. Number of Warehouses and Costs. Source: McKinnon, 2009, pg. 297.
18 Once again, the need to separate between different types of centralized and decentralized DCs is emphasized. In an order to produce a more accurate description, Rushton, et al. (2014) emphasized several other categories of DCs. These were defined as central (CDC), national (NDC), regional (RDC) or local DCs in ascending order in terms of number of warehouses per area. This approach was also previously adopted by Notteboom & Rodrigue (2010) using the US and Europe as examples with the following images:
As visible above, both local DCs and RDC have several DCs within the same market, however, the number of DCs vary substantially. While this approach adds a degree of complexity to the discussion, the general principles introduced previously still apply. The larger and fewer DCs there are, the more cost efficient and the more beneficial it is from a demand risk standpoint.
On the other hand, the more DCs there are the lower the supply risk is (Snyder & Shen, 2006.
Notteboom & Rodrigue, 2010. Rushton, et al. 2014. Schmitt, et al. 2015). As shown in Figure 1, there is a need to find an optimal number of warehouses to fit the needs of each company (McKinnon, 2009).
Figure 3. Local DCs and RDCs in Europe. Source: Notteboom & Rodrigue, 2010. Pg. 504.
Figure 4. Local DCs and RDC in USA. Source: Notteboom & Rodrigue, 2010 pg. 505.
19
4. Factors Impacting Location Selection & Number of DCs
In this chapter, a general framework detailing factors that impact warehouse location selection will be examined and applied in a European and Indian context.
4.1 Lumsden’s Framework
While most scholars agree that establishing a DC somewhere in between the customers and the supplier base is generally the most cost efficient there are a wide array of factors that affect the suitability of a location. As previously emphasized, one needs to analyze how the local conditions impact the supply and demand risk. However, which local conditions that should be considered must be determined in advance in order to stay true to the scope of the study. In his work, Lumsden (2006) analyzed several factors that impacts location selection. These are summarized in the image below:
Figure 5. Framework of Factors Impacting Warehouses and DC Location. Translated from Lumsden (2006) pg 613.
Factors Impacting Location of Warehouses and
DCs
Ability to Provide Service
Infrastructure Accessibility
Extended Market
Proximity to Competitors
Additional Factors
Proximity to Industrial Hub
or Major city
Ability to Secure Grants Ability to
Secure Loans Availability of
Buildings Tradition
Room for Expansion
Personal Contacts
Proximity to Markets
20 Though this framework may appear complicated and daunting at first sight, many of these factors are heavily correlated. For instance, room for expansion, proximity to markets/competitors, ability to provide service, and proximity to industrial hub or industrial hub or major city, are all factors which are strongly correlated. A DC cannot for instance, be located too far from either the end market where the customers are, or too far away from the production base. Being located far from both or either would increase transportation costs and diminish the DC’s ability to adequately provide the service it is meant to provide. Naturally, all these factors directly impact and are impacted by the land a company can acquire to build a DC on. As such, Land Acquisition serves as an excellent title for a category which can encompass these factors.
The ability to secure loans and ability to secure grants are both related to securing financing, which is a topic directly related to the ownership structure of a given DC. How a company decides to proceed with ownership of a particular plot of land or DC impacts how it can negotiate terms with the local government. A company may seek to lease their DC from the government rather than construct one on their own and therefore the availability of buildings that the local government has at hand and the personal contacts the company has are factors that impact the negotiating terms. This topic has also been explored by several authors such as De Koster & Warffemius (2005) who have studied the impact that ownership structures have on productivity and costs in warehouses and DCs.
A DC must also be located somewhere near infrastructure as without any supporting infrastructure, the DC can hardly provide the intended service as, once again, transportation costs would increase at an unsustainable rate. Infrastructure Accessibility already is a category that encompasses several factors, such as rail & rail terminals and highways. These have been studied in depth by several authors including, Cidell (2010) and Rodrgiue, Comtois & Slack (2006). Therefore, this category shall be included but will remain under a separate heading.
Tradition is primarily relevant when referring to IKEA’s European operations. One major reason why many of IKEA’s activities are conducted in Älmhult is due to the fact that IKEA was founded there. When referring to IKEA’s Indian operations, this factor is a moot point since IKEA has only recently entered the Indian market. Hence, this is not a topic that will be pursued in this thesis.
4.2 Issues of Land Acquisition
Though the choice of strategy relies extensively on advanced cost predictions and extensive calculations there are naturally, additional factors that impact the choice of strategy. In the
21 model introduced by Lumsden (2006), he stresses the importance of availability of buildings, room for expansion, ability to provide service and proximity to industrial hub or city. These factors are directly related to land acquisition. The availability of land thereby also directly impacts the strategy chosen. Depending on the local circumstances, there may be limitations to each strategy. While a large CDC will require a lot of consecutive land parcels, smaller DDCs will require several small parcels of land.
Contrary to what may be the most optimal location in terms of distances and other beneficial circumstances, many times the plot of land a company seeks may simply be unattainable. As Adams, Russell & Taylor‐Russell (1993) argued, one of the unique characteristics of land as a resource is that one cannot relocate it to wherever there is demand. As succinctly put by Mark Twain: “Buy land, they are not making it anymore,” (McIntyre, 2009). While this constraint may sound obvious, it adds several restrictions to how companies can operate in any given area.
Since most companies seek areas with similar attributes, such as accessibility to roads, there is likely to be stiff competition for land acquisition (Hesse, 2004. Notteboom & Rodrigue, 2010).
Any area with several competitors seeking land for development purposes will likely experience a surge in pricing. The type of DC that a company wants to develop has a direct impact on the amount of land required and in turn, the chances of acquiring such land. Whereas a local or regional DC can have a floor space ranging from anything in between 50,000 and 100,000 m2 many national DCs are likely to go well beyond those sizes (Hesse, 2004). Though such demands may not be an issue in countries where there is plenty of land available or where each parcel of land is rather large, this is certainly an issue in areas where land ownership is more fragmented.
In Europe, challenges related to land acquisition vary significantly between the different countries. In a report written by Shalak & Turk (2006), the regional differences for logistics development were analyzed and the following image was produced:
22 While some of the numbers presented above may be outdated, what is important to note is the relationship of land costs and country. As shown above in Figure 6, land costs appear to be lowest in France, Western Germany and Eastern Europe, while Italy, Spain, Central & Eastern Germany and the East coast of Sweden appear to have the highest land costs. Of course, the price of land is only one of the factors that impacts land acquisition and warehouse development. For instance, though land prices in Britain is one of the highest in Europe, Britain along with Germany and France account for approximately 65% of the total industrial/logistics investments in Europe with the Nordic countries like Sweden accounting for a sizeable part of the remaining investments (Mattarocci & Pekdemir, 2017). One must also bear in mind public concerns and unemployment rates. Countries or regions with comparatively high employment rates and with a population resistant to warehouse or DC construction are less likely to be accommodating towards companies attempting to expand their warehouse activities and can impact the price of land and a company’s attempt to acquire it (Shalak & Turk, 2006). While there are a significant number of companies in Europe that specialize in purchasing parcels of land in strategic areas and aggregating them into a single plot of land, the local governments still play an important role due to the fact that they control which areas can be designated under the appropriate zones (Raimbault, Andriankaja, & Paffoni, 2012. Mattarocci & Pekdemir, Figure 6 Land Prices in Europe. Source: Shalak & Turk, 2006. Pg. 46.
